This code is for the paper "Data Valuation without Training of a Model"

### Dependencies ###
Python 3.8.13
NumPy 1.22.3
PyTorch 1.12.1
Torchvision 0.13.1
CUDA 11.3.1 

### Code structure ###
cg_function.py : CG-score calculator using given dataset
                 You can import calc_cg_score and use the function to calculate CG-score in experiment
base_trainer.py : Base Trainer of network
- model		 Module of base_trainer
- utils		 Module of base_trainer

### How to run the code ###
python base_trainer.py --seed <SEED> --dataset <DATASET> --model <MODEL> --epochs <EPOCHS> --lr <LEARNINGRATE> --optim <OPTIMIZER> --scheduler <SCHEDULER> --regularizer <REGULARIZER> --batch_size <BATCHSIZE>